Peter Trudgill

Sociolinguistics and Linguistic Value Judgements: Correctness, Adequacy and Aesthetics

Series B: Applied and Interdisciplinary Papers ISSN 1435-6473 Essen: LAUD 1976 (2nd ed. with divergent page numbering 2013) Paper No. 10

Universität Duisburg-Essen

ii

 

Peter Trudgill

University of Bristol

Sociolinguistics and Linguistic Value Judgements: Correctness, Adequacy and Aestethics

Copyright by the author Reproduced by LAUD 1976 (2nd ed. with divergent page numbering 2013) Linguistic Agency Series B University of Duisburg-Essen Applied and Interdisciplinary Papers FB Geisteswissenschaften Paper No. 10 Universitätsstr. 12 D- 45117 Essen

Order LAUD-papers online: http://www.linse.uni-due.de/ Or contact: laud@uni-due.de

1

Peter Trudgill

Sociolinguistics and Linguistic Value Judgements: Correctness, Adequacy and Aesthetics

It is well-known that linguistics is a descriptive rather than perspective science, and that linguistics are concerned with describing and accounting for what speakers actually say rather than with what various “authorities” believe they ought to say.1 This means that linguistics are not normally prepared to say that some forms of language are “good” or “correct” and that other forms are “bas” or “wrong”. Some linguists (for example, Hall, 1950) have taken a much stronger and more proselytising stand on this issue than have others, but it is safe to say that the vast majority of linguists are agreed that notions of “correctness” have no part to play in objective discussions of language, at least as it is used by native speakers.

In this paper we want to argue that, at any rate in the English-speaking world, it is important, for educational reasons, for some linguists at least to make their views more widely known on the "correctness" issue. We also want to suggest that linguists should, in addition, resist value judgements about language on other counts, notably that of the "inadequacy" of certain language varieties, and that of the inferior "aesthetic" quality of certain types of speech, since judgements of this type are also important in the educational field. More particularly, we want to suggest that empirical sociolinguistic research that has been carried out both under experimental conditions and in the speech community itself can now be employed to demonstrate that value judgements of all three types are equally unsound. Section 1 will deal with "correctness" judgements; section 2 with notions of "adequacy"; and the rest of the paper will be concerned with various arguments connected with the "aesthetic" merits of different types of language.

1. Correctness in Language On the correctness issue, of course, linguists are set apart from most of the rest of the community, who are generally still persuaded that some grammatical forms and pronunciations are "wrong" and that it is possible for native speakers to make "mistakes" in this way. There are signs, it is true, that several decades of arguments from linguistics, together with other influences, are beginning to have some small effect. Many people involved in issues connected with language and education, for example, are starting to suggest - as most linguists have long believed - that what has traditionally been conceived of as a matter of "right" and "wrong" is in fact simply a question of dialect differences, and considerable discussion of topics of this sort is now taking place. This is especially true in the United States, where there is no heightened consciousness of dialect-related problems as

2

a result, largely, of educational difficulties connected with the racial situation. (See, amongst many others, Burling, 1973). There is still a long way to go, however, and many linguists addressing audiences of teachers and others have been very surprised to discover the death of feeling that any attack on the notion of correctness produces. This argument is an important one from an educational and social point of view. Many children, in many different language communities, are still discriminated against, both consciously and unconsciously, for using nonstandard dialects and low-status accents. This is particularly true in schools, but it also occurs in certain areas of employment, for example. Language, we can say, functions as a social disadvantage for these children - for no good linguistic reason.

It is therefore necessary for those linguists who are concerned with this problem to bring pressure to hear at appropriate points, and to marshal as many arguments as an be found to support their case. Simple assertions that no one dialect is any more “correct” than any other, such as are often found in introductory linguistics texts, are not adequate to persuade intelligent and educated laymen, who have long believed the opposite, of the justice of the linguist’s case. It is therefore fortunate that a body of data has now emerged from empirical sociolinguistic studies, particularly those into urban dialects, which can be used to demonstrate the true nature of value judgements based on correctness.

The best strategy for linguists to adopt would appear to be to point to the clear relationship between language and social class that emerges from these studies, and to suggest that judgements about "right" and "wrong" in language are not linguistic judgements at all, but social judgements. Most people, of course, are already aware of the correlation between social class and language, but they are most often conscious of it only in a rather unreflecting way. The advantage of the linguistic research that we are discussing is that it can illustrate the nature of this relationship in a clear, accurate and often somewhat dramatic way, the implications of which cannot be ignored. It is now possible, that is, to show very clearly that notions of "correctness", when applied to pronunciations and grammatical constructions used by native speakers, are based on social factors, and do not make linguistic sense.

A good example of the way in which this can be done is provided by an English grammatical construction which is universally considered to be ''wrong". This is the feature which is known to the English- speaking layman as "the double negative", and to linguists as multiple negation or negative concord. Most dialects of English permit constructions of the type:

I don't want none

He hadn't got no shoes

I can't find none nowhere

3

where indefinites in the sentence can be negated as well as the verb. The standard English dialect, on the other hand, permits negation of only one or the other, but not both. (Dialects vary, too, in the extend to which they permit multiple negation in different constructions (see Labov, 1972). Forms such as the following can be found in some dialects, but not in others:

She hadn't got hardly any

They stood there without no shoes on

We haven't got only one

It ain't no cat can't get in no coop)

In one of the earliest urban dialect surveys, that carried out in Detroit (Shuy et al 1967; Wolfram, 1969), multiple negation of the type illustrated in the first set of examples above was investigated, and the extent of its usage was correlated with social class membership. It emerged, in fact, that there was a very close relationship between the number of multiple negative forms used and the speaker's social class background. The average scores obtained by the different social class groups were:

upper middle-class: 2%

lower middle-class: 11%

upper working-class: 38%

lower working class: 70%

It is certain that comparable scores would be obtained from most other parts of the English-speaking world.

Another feature investigated in the Detroit survey was the usage of the -s marker on third-person singular present-tense verb forms. The same feature was also investigated in Norwich, England (Trudgill, 1974), where forms such as

He like it

and

It taste good

are also found. The percentage of 3rd-person –s used by members of different social-class groups in the two areas was: Detroit Norwich UMC 98 UMC 100 LMC 64 LMC 71 UWC 23 UWC 25 LWC 12 MWC 19 LWC 3

4

In another British study, in Reading, an opposite tendency has been investigated (Cheshire, forthcoming): The occurrence of the -s marker on other persons of the verb, as in I likes it, They goes every day. It has been noted that, whereas middle-class adults use 0% -s on persons other than third singular, lower working-class children employ a very much higher percentage. The implications of figures of this sort are clear, especially if they are allied to scores for other similar grammatical features studied in Detroit, New York (Labov et al, 1968); Washington (Fasold, 1972); Montreal (Sankoff & Cedergren, 1971); and elsewhere. Multiple negation; absence of 3rd-person singular -s; and presence of -s on other persons are all widely considered to be "wrong". The above figures show that they are most typical of working- or lower- class speech. We can interpret this to mean the following: grammatical forms which are most typical of working-class dialects have low status, because of their association with groups who have low prestige in our society. This low status leads to the belief that these forms are "bad" and they are therefore judged to be "wrong". Evaluations of this type are therefore clearly social judgements about the status of speakers who use particular forms, rather than objective linguistic judgements about the correctness of the forms themselves. (We cannot account for all "wrong" forms in this way. Many innovations, introduced as the result of linguistic change, may be considered "wrong" in a similar sort of way, regardless of which social class uses them. Social class dialect features, however, remain the principal source for "correctness" judgements).

With accurate information of this sort at their disposal, it should be possible for those linguists who are concerned to do so to begin to persuade the rest of the community that "correction" of "mistakes" is in reality the imposition of an alien dialect. This view is in fact gaining ground in some educational circles, where it is now recognised by many that attempts to eliminate non-standard dialects in school can have very harmful effects. (On the other hand, some Colleges of Education have now begun to argue that although, say, Standard English is no more "correct" than other varieties, it is nevertheless "appropriate" to certain situations and should be taught f r that reason. "Appropriateness" can easily become simply “correctness” under another name, and our view is that this approach should be treated with caution). Generally speaking, however, the debate seems to be moving in the direction of a discussion as to whether it is desirable to teach standard English to children in addition to their native dialects, or what are ultimately social reasons to do with employment prospects, and so on; or whether we should try to achieve greater "dialect tolerance'', and recognise that if children suffer because of attitudes to non-standard dialects, it is the attitudes that should be changed and not the dialects (see Fasold & Shuy, 1970).

In any case, we can say that linguists are generally agreed on the "correctness" issue, and that there are some signs that others, including, importantly, many in the world of education, are following their lead in attempting to avoid making value judgements of this

5

kind about language. Results from empirical linguistic research can perhaps help to speed this process.

2. Adequacy in Language The second area where value judgements are frequently made about language concerns the "adequacy" of particular types of language. The "correctness" issue is confined, for the most part to intra-linguistic comparisons (although of course the notion that 'Latin grammar was more "correct" than that of other languages has had an effect on a number of modern European languages- an example of this is the tendency to avoid sentence- final prepositions in formal styles of English).

On the adequacy issue, however, we have to take note both of inter-and intra-linguistic comparisons. As far as inter-linguistic comparisons are concerned, the majority of linguists appear to subscribe to the view that one language is as good and adequate as any other. As Halliday says: "all languages are equally capable of being developed for all purposes." (Halliday et al 1964, p. 100).

Many laymen, on the other hand, are convinced that English, for example, is inherently more adequate for certain (often academic) topics than, say, some indigenous African languages. This view, although it is often held by the speakers of the indigenous languages themselves, is surely without foundation, and there are educational and language-planning situations where it is important for linguists to say so.

There is also some discussion centering around the adequacy of pidgin and creole languages (see Labov, ms.). It is generally agreed that pidgins are not adequate for a number of purposes. When creolisation takes place, however, there is evidence that indicates that a number of developments occur which render the former pidgin entirely adequate for all the needs of its speakers in all its functions - although there are some who might be prepared to dispute this (see Craig, 1971). For a very interesting account of on-going creolisation processes, with evidence and examples of some of the devices employed to make a pidgin language fully adequate, see Mühlhauser (ms.).

From an educational point of view, however, the most serious issue would appear to be the intra-linguistic comparison of some varieties as more adequate than others. There are in fact two separate but related theories that can be interpreted as suggesting that some varieties of a language are less "adequate" than others. The first is Bernstein's theory of "elaborated" and "restricted codes". Bernstein (1962) writes, for example:

“The net effect of the constraint of a restricted code will be to depress potential linguistic ability”. And his earlier writings generally – which have been extremely influential amongst educationists – have been interpreted by many as indicating that “elaborated code” is in some ways a superior form of language. In his later writings, however, Bernstein (see the later papers in Bernstein, 1971) has stated that the one "code" is not superior to the other, although this is somewhat paradoxical in view of Statements of the

6

educational advantages of "elaborated code". (Bernstein says that "elaborated code" gives access to "universalistic orders of meaning" and that school "is necessarily concerned with the transmission and development of universalistic orders of meaning"). But it also becomes clear that Bernstein is now more concerned with the way in which speakers-language, rather than with the inherent adequacy of the "codes" themselves. He can therefore omit his work from this discussion. The second theory, which is based in part on Bernstein's writings as well as on misunderstandings of them, is the so-called "verbal deprivation hypothesis". This view holds - as it has been advanced by mainly American educationists and psychologists - that the language of some children is inadequate for certain purposes, such as handling abstract concepts and logical operations. Some writers on this topic make it plain that they regard some dialects (for example, American Black vernacular English) as inherently inferior in this way (Bereiter, 1966). Others relate the "inadequacy" less to specific features of dialect, and more to features of verbal interaction in working-class homes.

There is no space here to discuss the " verbal deprivation" or "language deficit" hypothesis in full. (This is done at greater length in Trudgill, forth-coming). But we should note here that linguists, once again, are for the most part strongly opposed to these ideas. Indeed, Labov (ms.) has written, in connection with the claim that some children are "verbally deprived": "The evidence put forward for this claim is transparently wrong, as linguists unanimously agree." And, again, empirical linguistic research can be employed to support the linguist's argument. In the first place, any linguist who has actually worked with working-class speakers and language knows form first-hand experience that working-class children are not "verbally deprived". But more importantly, data from linguistic studies can be used actually to demonstrate the strength of the linguist's case, as in Labov (1970).

It is, however, difficult to argue that there is no such thing as "verbal deprivation" except by cataloguing and illustrating at length and in detail the language resources and verbal skills of precisely those groups that are said to suffer from "language deficit". It is therefore fortunate that, to back up the convictions and the tape-recorded evidence of linguists who have carried out research in different speech communities, we have available at least one such carefully documented study, that of Labov and his colleagues carried out in Harlem, New York City (Labov, 1968).2

Once again, then, we can say that linguists reject the invidious comparison of linguistic varieties on the grounds of adequacy, and that, although in this case it is much more difficult to do so, we can provide data from sociolinguistic studies that can be used to argue this case. It is important that this case should be argued, since the educational implications of labelling some varieties of language as inherently inferior are very serious indeed. In particular, the setting up of "compensatory" educational programmes to “give” language to children who, of course, already have it, is both economically wasteful and psychologically and educationally dangerous. We can as linguists recognise that children

7

can be helped to develop expressive ability and verbal skills, but we cannot agree that some varieties of language are inherently more “adequate” than others.

3. Aesthetic Value and Language There is also a third area in which value judgements about language tend to be made: that of apparently aesthetic judgements about different languages, dialects and accents. This area appears to have received less attention from linguists than the other two, although in some respects it is equally important from an educational and social point of view. In fact, it appears, although there is little documentary evidence for this, that this is an area where many linguists are prepared, at least informally, to make as many value judgements as laymen. It seems to us that some linguists, at least, are just as prone as other people to say that, for example, Italian is more beautiful than Danish, - or vice versa. At the level of informal discussion, of course, views of this type have no particularly serious consequences. On the other hand, though, inter-linguistic comparisons may be harmful if, in multilingual situations, educational and other policy decisions are based on them.

At the intra-linguistic level, however, the discussion of the "aesthetic" merits of different varieties can be of some considerable importance - again particularly in the field of education. The fact is that, even if teachers are persuaded that all varieties of English or other languages are equally "correct" (and of course many of them are not persuaded of this) there is still a widespread feeling that some dialects and, in particular, some accents are much "nicer", "more pleasant" or "more beautiful" than others. It may appear at first sight that views of this sort are harmless and not worth the linguist's concern . However, the "aesthetic" argument is often used by teachers and others who attempt to change children's accents: there may be nothing "wrong" about the accent, the argument runs, and it may be perfectly comprehensible - but it is very "ugly"; it is therefore only fair to the children to give them the chance to speak in a more aesthetically satisfying manner. The grave danger here is that, whether views of this sort are accompanied by ridicule or by kindness, they lead speakers to disparage their own language, and children in particular to develop feelings of linguistic insecurity and even of what has been called "linguistic self-hatred". The result is, often, individuals who become, in certain circumstances, inarticulate and reluctant to express themselves. It would therefore be very useful if, by means of research data, we could demonstrate that aesthetic judgements, just like judgements concerned with correctness and adequacy, have no place in the objective evaluation of spoken language. (We are not, of course, concerned with literature). If this could be done, then perhaps something could be achieved towards persuading all speakers that their language is pleasant with, one would hope, a resultant increase in fluency and articulateness.

8

4. Inherent Value of Linguistic Varieties It may be that speakers of British English are unusual in the extent to which they are prepared to make value judgements on the aesthetic merits of linguistic varieties, but similar judgements are certainly made in other communities. Views of this sort were noted, for example, in research carried out in Norway (some of it reported in Trudgill, 1974a) as well as in work on Greek varieties of Albanian (Trudgill & Tzavaras, forthcoming). Similar evaluations have been noted with American English (Tucker & Lambert, 1969), Canadian French (d'Anglejan & Tucker, 1973) and Arabic (Ferguson, 1959).

There are two opposing explanations one can adopt for this widespread phenomenon. The first is a view that has been labelled by Giles et al (1974a) the "inherent value" hypothesis. This view maintains that some linguistic varieties are inherently more attractive and pleasant than others, and that these varieties have become accepted as standards or have acquired prestige simply because they are the most attractive. According to this view, for example, British R.P. is the most prestigious British accent because it has, as it were, risen to the top - or, according to Giles et al, been elevated to this position by a socially powerful group - as a result of it s inherent outstanding attractiveness.

It is certain that this hypothesis would find favour with a majority of laymen, including, crucially, many teachers. Terms such as "nicely spoken" and "with a pleasant voice", as used by the general public in England, are normally equivalent to "with an R.P. accent", and it is obvious that very many people simply take this hypothesis for granted. Some linguists, too, have accepted this position (see Wyld, 1934). There is also a certain amount of research evidence that could be interpreted as providing support for this hypothesis which we have to examine carefully before we can reject it altogether - as we would wish to be able to do in order to encourage linguistic self-confidence.

First, we have to note that a number of experiments carried out into listeners' subjective reactions to different linguistic varieties (Buck, 1968; Cheyne, 1970; Strengman & Noosley, 1967) have all succeeded in obtaining a pronounced uniformity of response. For example, Giles (1971a) has demonstrated that speakers with R.P. accents are more favourably evaluated on a number of different parameters than speakers with regional accents. They are almost universally evaluated as more intelligent, more reliable and more educated (Giles, 1971b) than other speakers. And there is also considerable evidence to show that most British people find R.P. the most aesthetically pleasing of all English accents (Giles, 1970). These experiments have all been carried out using the matched-guise technique where, although subjects believe they are evaluating different speakers, they are in fact reacting to the same speaker using different accents. In other words, different responses are entirely due to linguistic differences. The crucial point here is the high level of agreement that is obtained in the experimental results. Not only do subjects react to a change in accent, they all react in a very similar way. Supporters of the "inherent value"

9

hypothesis could therefore point to these facts and suggest that the majority response indicates that, for example, R.P. really is more beautiful than other accents.

In arguing against these views, however, we can suggest that the undoubted high uniformity of response is not due to any inherent aesthetic or other qualities. Rather it is the result of the fact that these reactions are due to certain cultural norms which are strong and pervasive, and which affect most listeners in a similar way. We could suggest, perhaps, that the subjects in these experiments have been "brain-washed" to an extent that renders objective responses on their part very unlikely. We shall discuss this point in section 5 below.

The second piece of evidence that we have to consider is rather similar. This consists of a series of overt statements made by informants, in a number of sociolinguistic studies, about different linguistic varieties. These statements, too, show a significant degree of agreement that could be interpreted as lending support to the “inherent value” hypothesis. It has emerged from a number of urban dialect studies, for instance, that even when speakers within a community themselves use very varied varieties of language, they nevertheless often appear to share, as a community as a whole, a common set of norms as to what is “good” and “bad” in the language. For example, Labov (1966) writes that “most New Yorkers think or feel that particular variants are better, or more correct, or are endowed with superior status” (p. 405). More importantly for our purposes, it appears that this level of agreement extends also to the apparently aesthetic. In the survey of Norwich reported in Trudgill (1974), for example, overt Statements such as the following were recorded, showing the high regard in which the aesthetic qualities of B.B.C. R.P. are held:

" I talk horrible, I think. But B.B.C. announcers and that, they really sound nice when they talk." house-wife, 45. "I think the Norwich accent is awful - but people you hear on the wireless, some of them have got really nice voices.'' night- watchman, 57.

Similarly, Macaulay & Trevelyan (1973) report comparable agreement from Glasgow as to the superior aesthetic qualities of "English" accents (the informants most probably had R.P. in mind):

"If you were an employer and somebody came in to see you in a broad Glasgow accent and then another man came in with an English accent, you'd be more inclined to give the English man the job because he had a nicer way of speaking." schoolboy , 15 . "There's no doubt the English - have us beaten there. Their speech is much preferable to ours.” commercial artist.

(Many other similar instances of agreement could be cited. American and British urban dialect studies, for instance, have shown that most speakers are prepared to praise prestige varieties as "pleasant", as well as, often to denigrate their own speech as "ugly"). Again we could consider this uniformity to be significant. If a majority of informants are willing to go

10

on record as saying that, say, R. P. is beautiful and their own accents are less so, can they all be wrong?

Against this argument we can cite, first, the cultural norms point we have already made above - if everyone has been subjected to the same cultural pressures, it is hardly surprising that they all produce the same sort of overt Statement. Secondly, we must also reckon with less overt feelings with speakers must also have. As Labov has said (1966, p. 108): "the socio-economic structure confers prestige on the middle-class pattern associated with more formal styles. [But] one can't avoid the implication that in New York City we must have an equal and opposing prestige for informal, working-class speech - a covert prestige enforcing this speech pattern. We must assume that people in New York City want to talk as they do, yet this fact is not at all obvious in any overt response that you can draw from interview subjects." Evidence for the existence of covert prestige of this type is provided in Trudgill (1972). Overt statements of the type we have cited, we can say, are by no means the whole story, since they are for public consumption only and take no account of more private or subconscious feelings.

The third point that we have to consider arises out of some work produced by Brown et al (1975). Their research has shown that French Canadian listeners can correctly allot French Canadian speakers to their social background on the basis of tape - recordings the speakers cade of a passage of prose. There is nothing particularly surprising a tout this, of course. What was more interesting that American listeners with no knowledge of French were able to do the some thing with a fair degree of success. This does not necessarily tell us anything about inherent aesthetic value, of course, but it does suggest that there may something inherent in, say, working- class accents which led the Americans to react in this way. In other words, we have to ask if this evidence can be used in any way to support the "inherent value" hypothesis.

We cannot argue conclusively against this, but we can at least shed serious doubts on the matter. One possibility that has to be considered, for instance, is that, since the matched guise technique was not employed in Brown’s work, listeners were reacting to social reacting to social - class - linked differences in reading and recitation skills. A second possibility is that there were perceptible differences of voice-quality between the middle- and working class speakers - something that is in fact suggested by comments made by the French Canadian listeners. It may well be that there is, as elsewhere (see Trudgill, 1974b, p. 185; Sachs, 1973), a relationship in Canadian French between sociological parameters and articulatory setting – long-term adjustments of the vocal tract which are acquired through social imitation, and are unconsciously and habitually maintained (Laver, 1968). If so, the American listeners may have been reacting to this. This does not, however, necessarily support the inherent value hypothesis, since it is not impossible (in fact linguistic area studies such as Emenau, 1956 suggest that it is quite likely that a similar relationship between the same paralinguistic features and social class exists in some varieties of

11

American English. If that were the case, it would provide an explanation for why the Americans reacted as they did.

We cannot assume, in other words, that certain types of setting are universally perceived as pleasant or unpleasant, regardless of culture or language. Indeed, there is some evidence to the contrary. Nasalisation, for example, is a component of setting commonly associated with many "unpleasant" Australian accents of English, but it is also a feature of many "nice" R.P. speakers (Laver, 1968). Similarly, pharyngealisation is a component of working-class Norwich voice-quality which middle-class Norwich people are almost entirely agreed is very unpleasant, (Trudgill, 1974b), but it is also a feature of some high status Arabic and German accents (see Honikman, 1964).

5. Imposed Norms and Linguistic Varieties We are not, then, in a position as yet to reject the "inherent value" hypothesis out of hand, but we have attempted to demonstrate that three types of evidence that could be cited in its favour also have other, perhaps more probable explanations.

Now we can attempt to argue more forcibly against the "inherent value" hypothesis by presenting some research data in favour of another, opposing hypothesis. Giles et al. (1974a) have labelled this competing view the “imposed norm” hypothesis. According to this view, different varieties of the same language are objectively as pleasant as each other, but are perceived positively or negatively because of particular cultural pressures operating in each language community. Standard dialects and prestige accents acquire their high status directly from the high status social groups that happen to speak them, and it is because of their high status that they are perceived as “good” and therefore as “pleasant”. Social pressures, Giles et al. argue, are placed on speakers to emulate ‘these varieties, and because of these pressures the varieties in question come to be regarded as desirable and superior on many counts, including the aesthetic. (I am aware that this might appear to be somewhat paradoxical, in that aesthetic judgements are possibly never entirely objective or culture free. The argument here, however, is that evaluations of language varieties, unlike those of, say, music, which are similarly culture bound, are the direct result of cultural pressures. It is true that we tend only to like and be in a position to judge music that is part of our own culture, but, although there is some broad degree of agreement as to that is good and bad in music, there is nothing at all like the striking total uniformity of response in relative evaluations of music that we find in evaluations of language).

This view, in some form or other, is one that would probably find favour with many linguists (see Spencer, 1958), even in the absence of research data. In this section, however, we present some research evidence which permits us actually to demonstrate, rather than simply assert, that at least partial validity of the “imposed norm” hypothesis.

In two empirical studies, of the experimental type, evidence has now been gathered which strongly suggests that the “imposed norm” hypothesis is worthy of acceptance in

12

many respects. The first study, reported in Giles et al (1974a), aimed to evaluate the merits of the two competing hypothesis. This it did by investigating the extent to which British adults with no knowledge of French were able to differentiate between varieties of French on aesthetic grounds, in a matched guise experiment. The value of this test was that, if speakers are evaluating varieties of a language which they are not familiar with and do not understand, they are unable to use their knowledge of cultural norms in order to formulate an aesthetic response, and they are similarly not subject to pressure from those cultural norms. We can be certain that they are reacting to the sounds of the accent, and to nothing else.

It is known (cf. d’Anglejan & Tucker, 1973) that, of the three varieties used in the experiment, French Canadians agree in evaluating educated European French as more pleasant than educated Canadian French, which in turn is rated as more pleasant than working-class Canadian French. The results of this experiment showed that the British listeners did not react in the same way. They were in fact in total disagreement with the French listeners as to the relative aesthetic merits of the three varieties. On average, all three were rated at approximately the same level of pleasantness. No significant differences at all in aesthetic evaluation emerged.

Because of the rather restricted nature of this data, it is fortunate that we also have results for another similar experiment that was carried out with a different group of subjects. In this case, the accents to be evaluated were matched guise recordings of Cretan and educated Athenian Greek (details of the experiment are given in Giles et al, 1974b). Greek informants had previously indicated that the Athenian variety was a prestige form of the language “possessing over the other varieties considerable advantages within that language community in terms of perceived pleasantness”. In the experiment, however, 46 British subjects with no knowledge of Greek showed no signs at all of any agreement on the relative aesthetic (and other) merits of the two types of Greek. In fact, the results show that the Cretan variety was rated as slightly more pleasant than the Athenian, although not significantly so. The results are given in Table 1.

Table 1: Mean ratings and t values for differences between Greek accents

Prestige Aesthetic Intelligent Sophisticated Athenian 4.70 5.39 4.22 4.89 Cretan 4.61 4.96 4.20 4.74 _________________________________________________________ t values 0.28 1.45 0.08 0.87 (d.f.=45) (The lower the mean rating, the more prestigeous and aesthetic the accent, and so on).

13

These results, although restricted to two languages and two experiments, do lend considerable support to the "imposed norm" hypothesis. They suggest that the uniformity of response obtained in the research discussed in section 1. is indeed the result of strong and pervasive cultural norms: remove these cultural norms, as with British listeners reacting to French and Greek, and the "aesthetic" response disappears too. So while French Canadians do not hesitate to evaluate French aesthetically, and in a uniform manner, British listeners do not concur with these evaluations, and react in a way that suggests that for them all varieties of French are equally pleasant.

6 . Social Connotations of Linguistic Varieties In the rest of this paper we want to argue that the "imposed norm" hypothesis is sound, as the experiments just reported suggest, but that it needs to be broadened and extended. The experiments demonstrated that aesthetic judgements about language varieties are culture-bound to a far greater extent than other types of aesthetic judgement. Apparently aesthetic responses to language, it seems, are in fact reactions to cultural norms. From the educational point of view , however, an important factor is that it can still be argued, by those hostile to the "imposed norm" hypothesis, that these judgements are worth acting on. It can be said, for instance, that just as Europeans are not trained to discriminate between good and bad Asian music, so British listeners cannot distinguish between beautiful and ugly French or Greek simply because they do not have sufficient experience of these languages. They are not part of the culture in question and are not therefore qualified judges. The results of our experiments can therefore be disregarded .

We therefore want to oppose this objection by suggesting that, while aesthetic norms in language certainly are imposed to an extent, aesthetic evaluations are not simply a matter of cultural norms. Rather, aesthetic judgements of linguistic varieties are the result of a complex of social connotations that these varieties have for particular listeners. (We use the term social here in its general sense. We do not intend to refer only to social class and status). Connotations of this type are by no means only a question of prestige or lack of it, and, crucially, they can and do vary within cultures. This means, first, that the “imposed norm” hypothesis is not entirely adequate; and, secondly, that we cannot argue that cultural outsiders are not qualified judges of aesthetic merit. Indeed, they are in a better position to make aesthetic judgements than most. They are unaware of the social connotations of the varieties involved, and are therefore reacting only to the sounds that they hear. The British listeners in the experiments reported in section 5. reacted as they did because they did not know that the social connotations of the different varieties wore. The fact that they rated all the varieties as approximately equal on aesthetic grounds, moreover, strongly suggests that there are good reasons for arguing that all dialects and accents are equally pleasant.

This hypothesis, which we can call the "social connotations" hypothesis, has already been argued for briefly by other linguists. Halliday (Halliday et al., 1964, p. 105) writes, for

14

example: "The chief factor in one's evaluation of varieties of a language is social conditioning." It suggests that it is not possible to obtain uniform responses from listeners on the aesthetic merits of different accents unless their social connotations are the same for all concerned; that these social connotations will not always be identical for all the members of a culture; and that if the social connotations of a variety are not known to a listener, he will not be capable of ranking it aesthetically relative to other varieties. Aesthetic judgements about language, that is, are just as much social judgements as those concerned with correctness.

(Note that we do not mean to imply that a listener is unable to make aesthetic judgements about an accent which he has never heard before. People clearly do this. We would suggest, however, that he must be familiar with the social connotations of at least one phonetic feature or combination of features of the accent. The English glottal stop realisation of /t/ would be a good example).

If we can show that this hypothesis is valid, we shall then he in a position to suggest that all varieties of a language are objectively equally pleasant. This would be helpful in many educational situations. What evidence then is there in favour of this "social connotations" hypothesis?

We can consider, first, an area where the "imposed norm" hypothesis is clearly inadequate. If we examine the aesthetic evaluations that are normally made in Britain of non - prestige accents, it is clear that, by and large, rural accents are regarded as aesthetically much more pleasing than urban accents by the vast majority of British people. As one Glaswegian said (Macaulay & Trevelyan, 1973): "it's the slovenly speech in the industrial areas I don't care for - these industrial cities, I don't like the accents they have." These accents which are most frequently singled out for opprobrium are the working-class accents of large cities such as Birmingham, Glasgow, Liverpool and London. On the other hand, rural accents such as those of the West Country and the Scottish Highlands are widely accepted as “beautiful”, "nice" or at worst "charming'' or "quaint". (One College of Education "speech" teacher distinguishes between "proper" accents such as "Yorkshire" and "Devon", and " unfortunate bastardised monstrosities" such as "Cockney" and "Birmingham").

We cannot attribute this phenomenon to the "imposed norm" hypothesis. No non- R.P accent can be regarded as a "norm" in England, since they all have depressed status relative to R.P. The fact remains, however, that low-status accents are evaluated differently, and that this differential evaluation has often no connection with any objective linguistic distance from the R.P. “norm”. Very often it is the urban accents which are objectively most like R.P.

We can therefore claim instead that this phenomenon is the result of the different social connotations rural and urban accents have for most British people. The vast majority of British people new live in towns, and many townspeople particularly the middle-class

15

who are probably instrumental in the formulation of mainstream public attitudes of this sort) have a romanticise nostalgic view of the countryside and the country way of life. They are much more realistic, on the other hand, in their assessment of the stresses and disadvantages of town life - particularly that of the urban working-class who se accents are most disliked. These views are transferred to the linguistic varieties associated with the different areas, which are therefore subject to different aesthetic evaluations. If, say, French listeners do not react in the same way, it is not because they know nothing about the English language as such. It is because they do not know a Birmingham accent when they hear one, and are not aware of its social connotations.

Secondly, it is possible to take this point somewhat further. The fact is that the social connotations of accents are not limited to a contrast between rural and urban. They are subject to much finer gradations. The "'social connotations" hypothesis is in fact particularly strongly supported by the way in which accents associated with even large urban areas which lack heavy industry and are widely regarded as pleasant are much more favourably evaluated than others. This is true, for instance, of the city of Bristol and the Bristol accent. Similarly, rural accents from less well-known or attractive areas such as parts of Lincolnshire or East Anglia are generally not rated as so attractive as accents from more stikingly beautiful parts of the country.

Perhaps, though, the most striking evidence in favour of this aspect of the social connotations hypothesis comes from the work of geographers. Although we have no figures for this, it is probable that there would be widespread agreement in England that the most unattractive accents in the country are those of the West Midlands in general and Birmingham in particular. In work reported in Gould & White (1974), research has been carried out into "mental maps" that people have of different areas. Studies of "environmental perception" have investigated subjects' images of Britain by asking them to plot an a map their preferences for where they would like to live if they had a free choice. By combining responses from different parts of the country, a "national perception surface" was constructed. This shows that the midlands, and the West Midlands in particular, is perceived as much less desirable than the surrounding areas in the country. We can suggest that it is not an accident that accents from this area regarded in the same way.

Thirdly, we can present evidence which shows that when social connotations vary, or information about them is lacking, then aesthetic responses to linguistic varieties also vary or disappear. In the Norwich survey, for instance, it emerged that working-class Norwich speech was consistently both distinguished from that of rural Norfolk (the area surrounding Norwich) and considered aesthetically inferior to it by local people. "Outsiders", however, - unless they had been in the city for some considerable time - did not rate Norwich speech as more unpleasant than-

16

Table 2: Correlation matrix – rank orderings on aesthetics parameter

England (S) Scotland Ireland (S) Canada (E) USA (E) USA (W)

(N) (23) (19) (28) (28) (22) (47)

England (S) - .82** .35 .16 .42 .63

Scotland - .35 .04 .06 .59

Ireland (S) - .59 .87** .41

Canada (E) - -70* .48

USA (E) - .39

USA (W) -

** = highly significant

*= significant

Cut of the fifteen pairings, there are only three significant agreements as to aesthetic merit. The first, not surprisingly, involves the two British groups, which comprised the only listeners to be well-informed as to the location and social connotations of the British accents. The second significant agreement concerns the U.S.A. (E) and Canadian groups. This may be the result of the relative geographical proximity of these two groups (tests were carried out in Virginia and Montreal) but we are very far from being able to show that this is so. And, finally, there is a high level of agreement between the U.S.A. (E) and Irish groups. We have no specific explanation for this. The crucial point, however, is that none of these three groups agree with the British ratings, and that there is no overall agreement.

The results show, then, that there is very little correlation indeed between the results from the different areas - and it should be noted that agreement would be even lower if we excluded the results for R.P., the social connotations of which were obviously known to all. British, American, Canadian and Irish listeners do not make the same aesthetic responses to British accents. Six different groups of listeners made very different aesthetic evaluations of the ten accents - and this in spite of the fact that this was not a matched guise experiment and their responses were therefore undoubtedly influenced by voice quality and reading.

We therefore feel justified in claiming that these results lend strong support to the social connotations hypothesis. The results from the different areas vary, we suggest, because the social connotations of the different accents either vary for different listeners or are not known. The information available to the Scottish and English listeners was not shared by the other groups, and they therefore reacted differently. And even where accents were recognised, it is clear that their social connotations varied from place to place. The "ugly" London accent, which was rated 10th by both the English and the Scots, was rated 3rd by the Canadians and eastern Americans, and 4th by the Irish. It is also worth noting that the major point of disagreement between the Scots and English groups - the fact that the Scots placed R.P. only second - is probably not unconnected with the different social connotations, involving nationalist sentiments, this accent has in the two countries.

17

The "inherent value" hypothesis, too, seems again to be disproved. If some of these accents were genuinely inherently more attractive than others, we would have expected a far higher degree of agreement than that portrayed in Table 2. It cannot be claimed, either, that the differential ratings are due to the fact that the judges were unqualified, since all were native speakers of English.

Conclusion We can conclude that those linguists who are concerned with problems of this nature should attempt to persuade others that the belief that some varieties of language are wrong and/or inadequate is untenable. We cannot, on the other hand, tell people that their aesthetic responses are false. Rather, we should encourage teachers and others - not to abandon their aesthetic judgements - cut to recognise them for what they are: the result of a complex of social, cultural, regional, political and personal associations and prejudices. Most listeners know of linguistic varieties that they do not like, but we should recognise that these feelings are very subjective and have no basis in objective linguistic fact. In particular, feelings of this sort should not be allowed to influence teachers' attitudes and policies towards children's language - the more so since they are likely to produce linguistic insecurity, and are in any case almost certainly not shared by all members of the wider culture. In the classroom there is a big and important difference between "Birmingham speech is ugly", a Statement of apparently objective fact, and "I personally find Birmingham speech unattractive" which, even if better left unsaid, is nevertheless a recognition of the subjectivity of responses cue to social connotations.

18

References D’Anglejan, A. & G. Tucker (1973). “Sociolinguistic Correlates of Speech Style in

Quebec” in R. Shuy & R. Fasold (eds.) Language Attitudes: Current Trends and Prospects (Washington: Georgetown U.P).

Bereiter, C. (1966). “An Academically Oriented Pre-School for Culturally Deprived Children”. In F. Hechinger (ed.). Pre-School Education Today (New York: Doubleday).

Bernstein, B. (1962). "Linguistic Codes, Hesitation Phenomena and Intelligence". Language and Speech 5, 31 -46.

Bernstein, B. (1971). Class, Codes and Control Vol. 1. (London; Routledge & Kegan Paul). Brown, P., W. Strong & A. Rencher (1975). “Acoustic Determinants of the Perceptions of

Personality from Speech”. IJSL (in press).

Buck, J. 1968. " The Effects of Negro and White Dialectal Variations upon Attitudes of College Students". Speech Monographs 35, 181 - 6.

Burling, R. (1973). English in Black and White (New York, Rinehart & Winston).

Cheshire, J. (forthcoming). Grammatical Variation in the English of Adolescents in Reading (Ph.D. the sis, Reading University).

Cheyne, W. (1970). "Stereotyped reactions to speakers with Scottish and English regional accents". B.J. of Soc. & Clinical Psychology 9, 77-9.

Craig, D. (1971). "Education & Creole English in the West Indies: some sociolinguistic factors". In: D. Hymes (ed.) Pidginisation & Creolisation of Languages (C.U.P.).

Emeneau, M. (1956). "India as a Linguistic Area" Language 32, 3-16.

Fasold, R. (1972). Tense Marking in Black English (Washington: Center for Applied Linguistics).

Fasold & P. Shuy (eds.) (1970). Teaching Standard English in the Inner City (Washington: Center for Applied Linguistics).

Ferguson, C. (1959). Diglossia". Word 15, 325 - 40. Giles, H. (1970). "Evaluative reactions to Accents" Educational Review 22, 211 - 27.

Giles, H. (1971a). "Patterns of evaluation in reactions to R.P., South Welsh and Somerset accented Speech". B.J. of Soc. & Clinical Psychology 10, 280 - 1.

Giles, H. (1971b). "Teachers' attitudes towards accent usage and change". Educational Review 24, 11- 25.

Giles, H., R. Bourhis & A. Davies (1974a). "Prestige Speech Styles: The Imposed Norm and Inherent Value Hypotheses". In W. ~McCormack & S. Wurm (eds.) Language in Anthropology IV: Language in Many Ways. (The Hague: Mouton).

19

Giles, H., R. Bourhis, P. Trudgill & A. Lewis (1974b). " The Imposed Norm Hypothesis: A Validation". O.J.S.

Gould, P. & R. White (1974). Mental Maps (London: Penguin). Hall, R. (1950). Leave Your Language Alone. (New York: Anchor).

Halliday, M., A. McIntosh & P. Strevens (1964). The Linguistic Sciences and Language Teaching. London: Longman).

Labov, W. (1966). The Social Statification of English in New York City. (Washington: CAL):

Labov, W. (1970). “The Logic of Non-Standard English” in J. Alatis (ed.). Georgetown Monographs 22, G.U.P.

Labov, W. ((1972). “Negative Attraction and Negative Concord in English Grammar”. Language 48, 773-818.

Labov, W. (ms.). “On the Adequacy of Natural Languages”. Labov, W., P. Cohen, C. Robins & J. Lewis (1968). A Study of the Non-Standard English

of Negro and Puerto Rican Speakers in New York City. (Philadelphia: U.S. Regional Survey).

Laver, J. (1968). "Voice Quality and Indexical Information", B.J. of Disorders of Communication 3, 43-54.

Macaulay, R. & G. Trevelyan (1973). Language, Education and Employment in Glasgow, (S.S.R.C. report).

Mühlhäusler, P. (ms.). The Creolization of New Guinea Pidgin, (R.S.P.S., Australian N.V., Canberra).

Sachs, J. et al (1973). "Anatomical & Cultural Determinants of Male & Female Speech" in R. Shuy & R. Fasold (eds.). Language Attitudes: Current Trends & Prospects (C.U.P.)

Sankoff, G. & H. Codergren (1971). "Some results of a sociolinguistic study of Montreal French". In R. Darnell (ed.) Linguistic Diversity in Canadian Society (Champaign, III.: Ling. Research Inc.).

Shuy, R., W. Wolfram & W. Riley (1967). A Study of Social Dialects in Detroit (Washington: Education Office).

Spencer, J. (1958). “RP-some problems of interpretation" Lingua 7, 7-29. Strongman, K. & J . Woosley (1967). "Stereotyped reactions to regional accents". B.J. of

Soc. & Clinical Psychology 6, 164 - 7.

Trudgill, P. (1972). "Sex, covert prestige & linguistic change in the urban British English of Norwich", Language in Society 1, 179 - 95.

20

Trudgill, P. (1974a). "Linguistic change & diffusion: description & explanation in sociolinguistic dialect geography" , Language in Society, 3.

Trudgill, P. (1974b). The Social Differentiation of English in Norwich (London: C.U.P.). Trudgill, P. (forthcoming). Accent, Dialect and the School (Arnold).

Trudgill, P. & G. Tzavaras (forthcoming). A Sociolinguistic Study of Albanian Dialects in the Athens Area of Greece. (S.S.R.C. report).

Tucker, R. & W. Lambert (1969). "White & Negro listeners' reactions to various American-English dialects". Social Forces 47, 463-8.

Wolfram, W. (1969). A Sociolinguistic Description of Detroit Negro Speech. (CAL).

Wyld, H. (1934). "The best English - a claim for the superiority of Received Standard English", S.P. E. Tract 39.